# \#bug

**URL:** https://openturns.discourse.group/tag/bug/24.md

[Latest](https://openturns.discourse.group/latest.md) · [Categories](https://openturns.discourse.group/categories.md) · [Tags](https://openturns.discourse.group/tags.md)

---

## [Covariance kernels for random processes with multidimensional output do not output symmetric matrices](https://openturns.discourse.group/t/covariance-kernels-for-random-processes-with-multidimensional-output-do-not-output-symmetric-matrices/55)

<div class="topic-metadata">

**Author:** [@josephmure](https://openturns.discourse.group/u/josephmure)\
**Replies:** 1\
**Last updated:** [November 23, 2020, 7:47pm UTC](https://openturns.discourse.group/t/covariance-kernels-for-random-processes-with-multidimensional-output-do-not-output-symmetric-matrices/55 "2020-11-23T19:47:17Z")

</div>

I think there is a mistake in the way covariance kernels for random processes with multidimensional output are handled by the library. In the following code block, I do not think operator() should return a CovarianceMat…